100% satisfaction guarantee Immediately available after payment Both online and in PDF No strings attached 4.2 TrustPilot
logo-home
Summary

Statistics 1 / Statistiek 1 - English Summary - VU Psychology Year 1

Rating
4.8
(5)
Sold
18
Pages
22
Uploaded on
01-12-2019
Written in
2019/2020

This is a visual summary based on the lectures of the Statistics 1 course for the English Psychology track at the VU. All concepts that are important for the exam and that have been discussed during the first 5 weeks are summarized and further visually explained using the examples given in the lectures.

Show more Read less
Institution
Module











Whoops! We can’t load your doc right now. Try again or contact support.

Written for

Institution
Study
Module

Document information

Uploaded on
December 1, 2019
Number of pages
22
Written in
2019/2020
Type
Summary

Subjects

Content preview

Statistics = a
body of methods for obtaining &
analyzing data


↳ provides methods for •
Design =
planning on how to gather data


Description =
summarizing data
-

D descriptive statistics


Inference =

making predictions ( for generalization ) based on the data -
D inferential statistics




Statistically = •


Probability often applies deduction -
o known ing the details of a
population ,
how likely is a certain


(sample) outcome? →
general to specific

Statistics often applies induction
( sample) what

given a cer tain outcome
,
can we
say about


the population & with what probability → specific to general



Similarities -

both work with randomness

-


Statistics is used to describe a
population
-

Some stat techniques first make assumptions about the

before her ) be true
population determining how
likely it is to

( Ho ,
HA ) ↳ based on falsification


Statistics methodology of how
you should perform empirical (pla
us
methodology systematic research
• =
=
.




way -
-





Statistics = the tools
-
needed to perform that empirical research




Week 1 : chapter 1 ,
2
,
3





Population total set of subjects of interest relevant for a research question (can be conceptual )

Patter the population ( e.g in %) usually inferred from statistic

=
numerical summary of .
a




Sample = a subset of that population on which the
study collects data .
the actual participants


Statistic = numerical
summary
of the sample leg religion
.


among the sample in %)
-
D a sample statistic often estimates a
population parameter (with a
margin of error )




Uariable-obseruedcharacteristicthatcanuaryamongsubj.ec





↳ can take on different forms

⑦ Types -0 behavioral ,
stimulus , subject & physiological variables




② Place on the measurement scale
discrete
categorical &
• Qualitative ( categorical ) → -0

Me -0 discrete


-0 Continuous or discrete
to

Quantitative ( numerical )



!




③ Range
• Discrete = measure unit is indivisible (siblings ) . . . . .





Continuous unit is divisible ( height)
-




=
measure

, The quality of an inferential statistic depends on how representative the sample is of the population
-0 so
you
need a random
sample taken from
your sampling frame f- list of all subjects in the
population )

Using random numbers ( =
computer generated selection )




Sampling methods

↳ simple Random sampling choosing random difficult
=
assigning everyone a number & numbers -0


'
↳ systematic sampling =
e.g .
Choosing every 4th person in a Room , using a skip number
'




↳ Cluster sampling =
choosing a few clusters within a
population leg . 100/360 high schools )
(strata )
↳ stratified Random sampling from
=
Selecting participants particular demographic categories in a




way that is proportionate to their membership of the population
↳ from
multistage sampling =
choosing a Random cluster ether
randomly selecting individuals it




What sample to Use depends on • the composition of the target population
-





the research question

the
feasibility to

obtain the sample




differeabetobseruedsmpksa.is#thepopuationparameercanbebecmsef :



1 .
Natural variation between samples ( is why we use a
margin 01 error )
2 .
Problems / mistakes with the sample
••
Sampling error = natural sampling variation


Sampling ( non probability sampling

bias = Selective sampling e.g .
Volunteer sampling)
,


or under
coverage ⇐ lacking representation of certain population groups )

Response bias = incorrect answering by respondents (e.g .

yea saying ) or bad
question wording
••
Non response bias Selective bc be refuse to
=
participation some con 't reached or participate
-




Descriptive Statistics methods


In describing data ,
3 dimensions are important





⑦ Central tendency (e.g .
mean
,
median ,
mode ) -0 the mean is not a
good central tendency when there a re
many
outliers !


⑦ Spread / dispersion / variability ( e.g .
Standard deviation ) -0 a mean can be similar for two curves but the spread can differ !


③ Position ( e.g .
on the axis ) -0
you can look at
quartiles or percentiles of interest




Descrietivestatistics.br#-



4,4%7 ) !
#
Categorical variables Quantitative variables




Jar
(
fi!saijivgelamfregyency
tin )
:p: not
buttons counts or % distribution stemmata
'




,



Central tendency measure Mode (Weighted ) average (mean ) median mode
fin
, ,



✓ = 1 -




Dispersion measure Variance ratio N
Range standard deviation
,
inter quartile range



/
,




! /
,




-
Position measure
percentile quartile
, ,
minimax ,
median
,
2 -
score CSP from
-
mean
#
entre.spreaitioioefgureboxpot-ocaseswil.tn



values > 3x IQR
Calculationsfortheboxplot
with values between i. s 3 x IQR / QR = Q3 -
Qi
cases
-








I lowest values no greater
lower limit = Q1 1.5 x IQR = lower wisher limit
highest
-




-0 extend to the
the ' QR
than -5 ×
Q3 IQR limit
'
* mean
wisher
upper limit = t 1.5 x
upper
=




line median
-
✓ To box = inter quartile range
thus 50%01 observations
-0 does not mean the wisher extends up to there
( top ,
-




but to the last nr .
within the limit

, skewed right skewed left




whatfigureretochoosedepadsono.tk
scale of the variable ( qualitative or quantitative )

Skewness of the distribution


Outliers in the data




Standard deviation s of n observations is



S=✓EnG
-
-



which means s=Fm%¥ts→ Because we first square each deviation & then sun those squares .




Sample size 1
It's
wrong to first add deviations
-




together & then square them

q reason for n - e is


Variance = S2
discussed in ch -
5




Week 2 : chapter 4 & 5



Probability The
=
probability of an outcome is the
proportion of times that outcome would occur in a
very long
So
long frequency
'
of it's relative
'


sequence observations -• a -
run distribution




Basicprobabilit-y.us
• P (A) -0 notation of probability of outcome A

p (not A) Pla ) that
Probability

= 1 -

-0 outcome A does not occur



PCA or B) = PH) t PCB ) -
D
probability of outcome A OR
Ag
B


P (A and B) = PCA ) x PCB gives A ) - D probability that booth A IB will occur when B is defeat on A


P (A and B) = Pla ) x PCB) →
probability that both A- & B will occur when both independent
-
-




Probability distribution = lists possible outcomes & their probabilities



£8 For discrete variables :
you assign a
probability for each possible value of the variable , using a p between o -


T


and everything together adding up to 7


e g -
.
ideal hr .
Of children




|#B For continuous variables : you assign probabilities to intervals of numbers -
b
you then can tell the
probability that

the
a variable will fall in a particular interval using the areas of probability under curve

, teare3typesofdistributions(ofprobabilit#

⑦ The
population distribution statement of the frequency with the
=
a which units of make
analysis up a
population
are Expected to be) observed in the various categories that make up a variable


-8 often unknown


TBA parameters : M mean


o standard deviation


N population size




② The sample distribution = a statement of the frequency with which the units 01
analysis make up a
sample
-




are Expected to be) observed in the various categories that make up a variable


-
Bo should look similar to the population distribution

poor statistics : I mean


s standard deviation


sample size

\
n




③ The sampling distribution = a statement oh the frequency with which values of statistic s are (expected to be ) observed

when a number of random samples are drawn from a
given population
BB specifies the probabilities for the possible values the statistic can take (due to natural
variation)
to describes statistic across samples :
MJ mean
,
will equal M (or tested )
standard deviation Standard
og =
error

sampling

D infinite samples of size n

population
y
Centimeter if you take sufficiently large samples from the population with
c-




£TdTFerds on
sample
replacement then the sampling distribution of
sample means will be size



approximately a normal distribution

TB We
generally view the mean of the
sampling distribution as the
population mean so Mg =
M


IB The standard deviation be the standard sample
of the
sampling distribution can seen as error of drawing a



from that particular population so : og = Fin -0 aka . dependent on size of sample taker
,




bigger sample = smaller standard er ror



Toda No the distribution the be
matter the shape of
population , sampling distribution will
normally distributed .




This normality property is used for significance & constructing confidence
testing intervals


to
Karge sample becomes more important when the population distribution is
relatively skewed (for validity )
£6.34
Get access to the full document:
Purchased by 18 students

100% satisfaction guarantee
Immediately available after payment
Both online and in PDF
No strings attached


Also available in package deal

Reviews from verified buyers

Showing all 5 reviews
1 year ago

3 year ago

4 year ago

6 year ago

6 year ago

4.8

5 reviews

5
4
4
1
3
0
2
0
1
0
Trustworthy reviews on Stuvia

All reviews are made by real Stuvia users after verified purchases.

Get to know the seller

Seller avatar
Reputation scores are based on the amount of documents a seller has sold for a fee and the reviews they have received for those documents. There are three levels: Bronze, Silver and Gold. The better the reputation, the more your can rely on the quality of the sellers work.
FabienneDenberg Vrije Universiteit Amsterdam
Follow You need to be logged in order to follow users or courses
Sold
540
Member since
6 year
Number of followers
242
Documents
0
Last sold
2 days ago

VU psychology honors student summarizing subjects on the Ipad Pro. I upload summaries for each subject at the latest a week before the exam week.

4.6

118 reviews

5
76
4
36
3
3
2
1
1
2

Recently viewed by you

Why students choose Stuvia

Created by fellow students, verified by reviews

Quality you can trust: written by students who passed their exams and reviewed by others who've used these revision notes.

Didn't get what you expected? Choose another document

No problem! You can straightaway pick a different document that better suits what you're after.

Pay as you like, start learning straight away

No subscription, no commitments. Pay the way you're used to via credit card and download your PDF document instantly.

Student with book image

“Bought, downloaded, and smashed it. It really can be that simple.”

Alisha Student

Frequently asked questions